Temporal trends and regional clusters of tuberculosis-HIV coinfection outcomes in Brazil.

OBJECTIVE

To investigate the temporal trends and regional clusters of tuberculosis treatment outcomes among people with HIV in Brazil, as well as their relationship with socioeconomic and programmatic indicators.

METHODS

Ecological study with data from people living with HIV who initiated and completed tuberculosis treatment from 2015 to 2021 in Brazil. We described the semiannual trends of cure, treatment interruption, and death rates in Brazilian states using joinpoint regression models. Cluster analyses stratified by three population sizes were conducted using the k-means method to identify clusters in 510 immediate geographic regions. Socioeconomic and programmatic indicators related to treatment interruption and death were included in multivariate negative binomial regression models.

RESULTS

A total of 54,362 tuberculosis treatments in people with HIV were analyzed, with cure rates of 55.51%, interruption rates of 23.33%, and death rates of 21.16% for the period. Nationally, cure rates remained stable, while treatment interruption increased by 2.54% per semester (ranging from 1.59% to 3.70%) and death increased by 9.31% (ranging from 7.41% to 17.24%). The states of Ceará and Amapá showed the worst trends for treatment interruption and death. Regions with greater income inequality, higher percentages of workers with only primary education, higher household density, and broader private health coverage had higher rates of treatment interruption and death. Conversely, regions with higher expected years of schooling and hospitalizations for primary care-sensitive conditions had lower probabilities of these outcomes.

CONCLUSION

Nationally, despite stable cure rates, tuberculosis treatment interruption and death among people with HIV increased. Regional disparities in the relationship between socioeconomic and programmatic indicators and outcomes suggest inequities in access to and adherence to tuberculosis treatment across Brazil.